Testing Hereditary Properties of NonExpanding BoundedDegree Graphs


 Mervin Melton
 3 years ago
 Views:
Transcription
1 Testing Hereditary Properties of NonExpanding BoundedDegree Graphs Artur Czumaj Asaf Shapira Christian Sohler Abstract We study graph properties which are testable for bounded degree graphs in time independent of the input size. Our goal is to distinguish between graphs having a predetermined graph property and graphs that are far from every graph having that property. It is believed that almost all, even very simple graph properties require a large complexity to be tested for arbitrary (bounded degree) graphs. Therefore in this paper we focus our attention on testing graph properties for special classes of graphs. We call a graph family nonexpanding if every graph in this family has a weak expansion (its expansion is O(1/ log 2 n), where n is the graph size). A graph family is hereditary if it is closed under vertex removal. Similarly, a graph property is hereditary if it is closed under vertex removal. We call a graph property Π to be testable for a graph family F if for every graph G F, in time independent of the size of G we can distinguish between the case when G satisfies property Π and when it is far from every graph satisfying property Π. In this paper we prove that in the bounded degree graph model, any hereditary property is testable if the input graph belongs to a hereditary and nonexpanding family of graphs. As an application, our result implies that, for example, any hereditary property (e.g., kcolorability, Hfreeness, etc.) is testable in the bounded degree graph model for planar graphs, graphs with bounded genus, interval graphs, etc. No such results have been known before, and prior to our work, very few graph properties have been known to be testable for general graph classes in the bounded degree graph model. A preliminary version of this paper, entitled On Testable Properties in Bounded Degree Graphs, authored by the first and third authors, appeared in the Proc. of 18 th Symposium on Discrete Algorithm (SODA), New Orleans, Louisiana, 2007, Research supported in part by NSF ITR grant CCR , Centre for Discrete Mathematics and its Applications (DIMAP) and EPSRC grant EP/D063191/1, NSF DMS grant , and DFG grant Me 872/83. Department of Computer Science, University of Warwick, Coventry, CV4 7AL, U.K. Microsoft Research. Department of Computer Science, University of Paderborn, Paderborn, Germany. Work done while the author was visiting Rutgers University. 1
2 1 Introduction The area of Property Testing deals with the problem of distinguishing between two cases: that an input object (for example, a graph, a function, or a point set) satisfies a certain predetermined property (for example, being bipartite, monotone, or in convex position) or is far from satisfying the property. Loosely speaking, an object is ɛfar from having a property Π, if it differs in an ɛfraction of its description from any object having the property Π. For example, when the object is a (dense) graph represented by an adjacency matrix and the property is bipartiteness, then a graph is ɛfar from bipartite if one has to delete more than ɛ n 2 edges to make it bipartite. Given oracle access to the object, many objects and properties are known to have randomized property testing algorithms whose time complexity is sublinear in the input description size; often, we can even achieve running time completely independent of the input size. In particular, sublineartime property testing algorithms have been considered for graphs and hypergraphs, functions, point sets, formal languages, and many other structures (for the references, see the excellent surveys [14, 16, 17, 23, 29]). After a series of results for specific problems, recently much attention has been devoted to study a more general question: which properties can be tested in time independent of the input size. This question has been especially extensively investigated for properties of dense graphs represented by an adjacency matrix. It turned out that property testing in dense graphs is closely related to Szemerédi s regularity lemma. Very recently, this relation has been made explicit by showing that any property is testable if and only if it can be reduced to testing the property of satisfying a finite number of Szemerédipartitions (see [2]). Furthermore, it has been shown that a property is testable with onesided error if and only if it is either hereditary or it is close (in some welldefined sense) to a hereditary property (see [6] and [11, 26]). While property testing in dense graphs is relatively wellunderstood, surprisingly little is known about property testing in sparse graphs. Properties of sparse graphs are traditionally studied in the model of bounded degree graphs introduced by Goldreich and Ron [21]. In this model, the input graph G is represented by its adjacency list and the vertex degrees are bounded by a constant d independent of the number of vertices of G (denoted by n). A testing algorithm has a constanttime access to any entry in the adjacency list by making a query to the i th neighbor of a given vertex v, and the number of accesses to the adjacency list is the query complexity of the tester. A property testing algorithm is an algorithm that for a given graph G determines if it satisfies a predetermined property Π or it is ɛfar from property Π; a graph G is ɛfar from property Π if one has to modify more than ɛ d n edges in G to obtain a graph having property Π. These results, imply that in the adjacency matrix model, essentially any natural graph property can be tested with a constant number of queries. Unlike the adjacency matrix model, in the bounded degree graph model only a few, very simple graph properties (like connectivity) are known to be testable in constant time [21] and the main research focused on designing property testers with sublinear query complexity (like, O( n) tester for bipartiteness [21]). Even more, it has been demonstrated that unlike in the adjacency matrix model, in the bounded degree model many basic properties have a nonconstant query complexity. For example, acyclicity in directed graphs has Ω(n 1/3 ) query complexity [9], the property of being bipartite has query complexity Ω( n) [21], and the query complexity of testing 3colorability is Ω(n) [10]. In fact, it is believed that very few properties can be tested in the bounded degree model with o( n) or even with o(n) query complexity. In this paper, we take a new approach and we study property testing in the bounded degree model under the assumption that the input graph belongs to a certain (natural) family of graphs. The goal of this investigation is to identify natural families of graphs, such as planar graphs, for which many properties can be efficiently under the assumption that the input graph belongs to the family, even though the testing 1
3 problem may be very hard in the general case. For the rest of this paper, we say that a graph property is testable if it can be tested in time independent of the size of the input graph. A family of graphs is called nonexpanding if it does not contain graphs with expansion larger than 1/ log 2 n; (this is informally equivalent to the families of graphs with some good separator properties). A family of graphs is called hereditary if it is closed under vertex removal. Similarly, a graph property is called hereditary if it is closed under vertex removal. We show the following result: In the bounded degree graph model, any hereditary property is testable if the input graph belongs to a hereditary and nonexpanding family of graphs. The reader is referred to Theorem 1 for the precise statement of our main result. Hereditary graph properties have been extensively investigated in combinatorics, graph theory, and theoretical computer science (see also the recent results about testability of hereditary graph properties in the dense graph model [6]). The class of hereditary graph properties contains also trivially all monotone graphs properties (properties closed under removal of edges and vertices). Many interesting graph properties are hereditary, for example, being acyclic, stable (independent set), planar, perfect, bipartite, kcolorable, chordal, perfect, interval, permutation, having no induced subgraph H, etc. (see also [16, 28]). Our result implies that these properties can be tested (in the bounded degree graph model) when the input graph belongs to a family of graphs which is hereditary and nonexpanding. Examples of natural hereditary nonexpanding families are planar graphs, graphs with bounded genus, graphs with forbidden minors, unit disk graphs, interval graphs, (planar) geometric intersection graphs, etc. We are not aware of any prior results showing testability of these properties for nontrivial classes of graphs. 2 Preliminaries Let G = (V, E) be an undirected graph with n vertices and maximum degree at most d. Without loss of generality, we assume that V = {1,..., n}. We write [n] := {1,..., n}. Given a subset S V of vertices, we use G S = (S, E S ) to denote the subgraph induced by S, where E S = {(u, v) E (S S)}. We assume that G is stored in the adjacency list model for bounded degree graphs with maximum degree d. In this model, we have constant time access to a function f G : [n] [d] [n] {+}, such that f G (v, i) denotes the i th neighbor of v or a special symbol + in the case that v has less than i neighbors. Definition 2.1 A graph G is ɛfar from a property Π if one has to modify more than ɛdn entries in f G to obtain a graph with property Π. 2.1 Testing a property in a graph family In this paper, our main focus is on testing various graph properties for bounded degree graphs from certain graph families (e.g., planar graphs or unit disk graphs). An algorithm that is given n and has access to f G is called an ɛtester for a graph family F if it (a) Accepts with probability at least 2 3 any graph G F that has property Π. (b) Rejects with probability at least 2 3 any graph G F that is ɛfar from Π. If the ɛtester always accepts any graph G F that has property Π, then it is said to have onesided error. The ɛtesters presented in this paper have onesided error. They will in fact accept with probability 1 any graph that satisfies Π (even if it does not belong to F). 2
4 A property is called testable for a family F if for any fixed 0 < ɛ < 1 there is an ɛtester for F whose total number of queries to the function f G is bounded from above by a function, which depends only on ɛ and not on the size n of the input graph. Following [5], we define a property Π to be uniformly testable if there is an ɛtester for Π that receives ɛ as part of the input. A property Π is said to be nonuniformly testable if for every fixed ɛ, 0 < ɛ < 1, there is an ɛtester that can distinguish between graphs that have property Π from those ɛfar from having Π (which may not work properly for other values of ɛ). For a pair of disjoint vertex sets V 1, V 2 we denote by e(v 1, V 2 ) the number of edges connecting vertices from V 1 with vertices from V 2. For each vertex v V, we denote its neighborhood by N (v) = {u V : (v, u) E}. We generalize this notion to sets by defining N (S) = v S N (v) \ S. Furthermore, we let D(v, r) denote the set of vertices which have distance at most r from v, i.e., D(v, 0) = v, D(v, 1) = {v} N (v), etc. A graph G = (V, E) is called a λexpander, if for all S V with S n/2, we have N (S) λ S. With this, we can now define nonexpanding graph families. Definition 2.2 A family of graphs F is called nonexpanding if there exists a constant n F such that all graphs in F of size at least n F are not (1/ log 2 n)expanders Hereditary and nonexpanding graph families A family F of graphs is called hereditary if it is closed under vertex removal. Similarly, a graph property is called hereditary if it is closed under vertex removal 2. There are many interesting classes of families of graphs that are hereditary and nonexpanding. For example, the family of planar graphs is trivially hereditary, and also the classical planar separator theorem [25] implies immediately that it is nonexpanding. Indeed, the planar separator theorem implies that every planar graph with n vertices (for a sufficiently large n) has a subset of vertices A, 1 3 n A 1 2 n, such that N (A) O( n). Therefore, every planar graph with n vertices (n n 0 for some constant n 0 ) is not an O(1/ n)expander, and hence the family of planar graphs is nonexpanding. As the example of planar graphs shows, if a family of graphs has a good separator then it is nonexpanding. Therefore, all graph families with good separator properties (for graphs of bounded degree) are nonexpanding. Hence, other families of graphs (of bounded degree) that are hereditary and nonexpanding include, among others: the class of graphs with bounded genus, graphs with forbidden minor, interval graphs, etc. For example, the result for graphs of bounded genus and graphs with forbidden minor follow directly from the separator theorem such graphs. And so, Gilbert et al. [15] proved that any graph on n vertices with genus g has a separator of order O( gn), and Alon et al. [4] showed a similar results for graphs with forbidden minors: if G has no minor isomorphic to a given hvertex graph H, then G has a separator of size O(h 3/2 n 1/2 ). 3 Proof of the Main Result In this section we prove our main result by showing that the following algorithm is an ɛtester for any hereditary property Π and any hereditary nonexpanding family of graphs F. 1 The choice of the factor 1/ log 2 n can be relaxed. In fact, using known bounds one can replace 1/ log 2 n with 1/(log n log 2 log n). 2 There is, of course, no difference between a graph property and a family of graphs. We use the different terms in order to distinguish between the property we want to test and the family of graphs to which the input is assumed to belong to. 3
5 ɛtester(g, n, Π) sample a set S of s 1 vertices uniformly at random for each v S do U v = D(v, s 2 ) U = v S U v if G U does not satisfy property Π then reject else accept Clearly, the number of queries to f G is upper bounded by 2 s 1 d s 2, which for s 1 and s 2 being constants independent of n, gives the number of queries to be independent of n. We will give the exact values for s 1 and s 2, which are independent of n but do depend on ɛ and Π, at the end of our analysis, in the proof of Theorem 1. Since Π is hereditary, we know that our algorithm accepts any graph that has property Π (even if it does not belong to F). Thus, we only have to show that any graph that is ɛfar from Π and belongs to F is rejected with probability at least 2 3. We begin our analysis with the following lemma. Lemma 3.1 Let F be any hereditary nonexpanding family of graphs and let n F be the constant from Definition 2.2. Let δ be an arbitrary positive parameter. If G = (V, E) F satisfies n = V max{2n F, 2 2/δ2 } then one can partition V into two sets V 1 and V 2, such that V 1, V 2 n 4 and e(v 1, V 2 ) δ d n/ log 1.5 n. Proof : Since F is nonexpanding, every graph G F on n n F vertices is not a 1/ log 2 nexpander. Therefore, there exists a set S V of cardinality at most n 2 such that N (S) S / log2 n. We first observe that if S n 4, then we can take V 1 = S and V 2 = V \ S. Indeed, since N (S) S / log 2 n, there are at most dn/ log 2 n edges between V 1 and V 2. Therefore, if in addition n > 2 2/δ2, we can infer that e(v 1, V 2 ) dn/ log 2 n δ d n/ log 1.5 n, as needed. Assume then that S < n 4 and consider the graph G V \S (the induced graph on V \ S). Since F is hereditary, G V \S F, and V \ S > n F (recall that n > 2n F ), we can apply the same arguments as above to conclude that there is a set S (V \ S) of cardinality at most n 2 such that N (S ) 2 S / log 2 n. If we have S S n 4 then using the same arguments as above, we are done by setting V 1 = S S and V 2 = V \ V 1. Otherwise, we can replace S by S S and continue in the same manner. Eventually, we have a set S S with more than n 4 vertices and N (S S ) 2 S S / log 2 n. If we set V 1 = S S and V 2 = V \ V 1, then these sets will satisfy the condition in the lemma. Let us call a connected component nontrivial if it has more than a single vertex. The following is a corollary of Lemma 3.1. Corollary 3.2 For every hereditary and nonexpanding family of graphs F, there exists a positive constant c = c F, such that one can remove from any graphs G F a set of at most ɛ d n/2 edges, such that (i) Their removal partitions G into connected components C 1, C 2,... of size at most 2 c/ɛ2. (ii) Each connected component C i is an induced subgraph of G. 4
6 (iii) No edge connects in G two nontrivial connected components C i and C j. Proof : Let n F be the constant associated with F as in Definition 2.2, let G be any graph in F, and let δ be a parameter to be chosen later. We apply Lemma 3.1 to obtain two sets V 1 and V 2 with at most δ d n/ log 1.5 n edges connecting V 1 and V 2. Assume V 1 V 2 and let U = N (V 1 ). Since the number of edges between V 1 and V \ V 1 is at most δ d n/ log 1.5 n, we also have U δ d n/ log 1.5 n. Remove from G all edges incident to U. Since U δ d n/ log 1.5 n and G has maximum degree at most d, we removed at most δ d 2 n/ log 1.5 n edges from G. Next, let U 1 = V 1 and U 2 = V 2 \ U. Observe that for δ log 1.5 n/(4d) we have n 4 U 1, U 2 3n 4 and that there is no edge in G between U 1 and U 2. Then we recursively apply Lemma 3.1 on the induced subgraphs G U1 and G U2 ; we proceed recursively until we obtain a subgraph of size at most max{2n F, 2 2/δ2 }. In this way, we removed some number of edges from G and obtained a subgraph of G denoted H, on V (G) with connected components C 1,..., C q. Observe that the sets U obtained in the recursive calls will always result in trivial connected components, because we removed all edges incident to the vertices in U. Let H 1,..., H k be nontrivial connected components in our new graph. By definition, every C i has size C i max{2n F, 2 2/δ2 }. Similarly, our construction ensures that no edge is removed between any pair of vertices in a single H i and that there is no edge in G between any pair of graphs H i and H j. We now estimate the number of edges removed. By Lemma 3.1, the number of edges removed from G is upper bounded by function Q(n) defined by the following recurrence: { 0 if n max{2nf, 2 2/δ2 } Q(n) = δ d 2 n/ log 1.5 n + max 1 4 τ 3 {Q(τ n) + Q((1 τ) n)} if n > max{2n F, 2 2/δ2 }. 4 Since Q(n) = Θ(δ d 2 n), we can conclude that the graph H is obtained from G by removal of at most c δ d 2 n edges, for some absolute positive constant c. This yields the proof by setting δ = ɛ/(2dc ). Finally, recall that all the connected components of H had size C i max{2n F, 2 2/δ2 } 2 c/ɛ2 if we take c = c F = 2 d c n F. Let us explain the importance of the three properties of the resulting graph stated in Corollary 3.2. Property (i) ensures that every connected component is small. Property (ii) ensures that if we have some induced subgraph of a nontrivial connected component H i then it is also an induced subgraph of G. Property (iii) ensures that if we have a set of induced subgraphs Q i1, Q i2,..., Q il of graphs H i1, H i2,..., H il, then these copies of the subgraphs do not intersect in H. Therefore, if we define a graph Q with l connected components, where the j th connected of Q is isomorphic with Q ij, then Q is also an induced subgraph of G. 3.1 Hereditary graph properties It is well known (and easy to see) that any hereditary graph property Π can be characterized by a (possibly infinite) set of minimal forbidden induced subgraphs (see, e.g., [6, Section 4]). Let us denote by Hforb Π a minimal family of forbidden subgraphs for property Π. Notice that in general, Hforb Π may be an infinite family of forbidden graphs. Observe that, for example, if Π is the property of being bipartite, then Hforb Π can be chosen to be the set of all odd cycles, and if Π is the property of being chordal, then Hforb Π is the set of all cycles of length at least 4. For simplicity of presentation (but without loss of generality) we will assume that the graphs in Hforb Π contain no isolated vertices. The reason why we can make such an assumption is that every large enough 5
7 bounded degree graph G will always have an arbitrary large induced subgraph that consists of isolated vertices only. Therefore, in such cases, all large enough graphs will not satisfy H Π forb, and thus testing HΠ forb becomes trivial. Next, let us consider an arbitrary graph G F that is ɛfar from Π. By Corollary 3.2, we can remove from G at most ɛ d n/2 edges to obtain a graph H on the same vertex set for which each connected component has at most r = 2 c/ɛ2 vertices. Furthermore, if H 1,..., H k are the nontrivial connected components of H, then there is no edge in G that connects any of these connected components and each H i is an induced subgraph of G. Since G is ɛfar from Π, H is still ɛ/2far from Π. Since all connected components in H have size at most r (which is independent of n), H cannot contain as a subgraph any graph that has a connected component with more than r vertices. Let J r denote the family of all graphs whose connected components have size at most r (notice that J r is independent of G). We conclude that it suffices to consider the subgraphs in H Π forb J r. Corollary 3.3 If G F is ɛfar from Π, then H (define above) contains as an induced subgraph a graph from Hforb Π J r. The same holds if we remove from H any set of at most ɛ d n/2 edges. Let us denote by c(r) the number of connected (unlabeled) graphs on a set of at most r vertices; clearly c(r) 2 (r 2). Let us enumerate all possible connected graphs with at most r vertices by G1,..., G c(r). Then, we can define any graph G in Hforb Π J r as a c(r)ary integer vector f = f 1,..., f c(r), where f i denotes the number of copies of graph G i occurring as a connected component in G. In what follows, we call f a characteristic vector of G (with respect to Hforb Π and J r). Similarly, let us define a c(r)ary integer vector g H = g H 1,..., g H c(r) with g H i being the number of induced copies of graph G i in H. Notice the fundamental difference between the ways of counting copies of G i in G and in H: all copies of graphs G 1,..., G c(r) counted in the characteristic vector of G are disjoint while the induced copies of these graphs counted in g H can intersect. Lemma 3.4 Let F be a fixed hereditary nonexpanding family of graphs and let Π be a fixed hereditary property. Suppose that G F is a graph of degree at most d that is ɛfar from Π. Assume that we apply Corollary 3.2 on G and obtain a subgraph of G denoted H with the property that all connected components of H are of size at most r. Then, there exists a graph G Hforb Π J r with characteristic vector f = f 1,..., f c(r) such that for all 1 i c(r) it holds that if f i > 0 then g H i γ n, where γ = ɛ d/2 r2. Proof : Let G 1,..., G c(r) be all connected graphs of size at most r. We will first construct a graph H by removing some edges from H so that for any graph G i either H contains no copy of G i or it contains at least γ n such copies. We proceed sequentially over the graphs G 1,..., G c(r). For each G i we do the following: if the number of induced copies in the current graph obtained from H is smaller than γ n, then we remove all the edges of any connected component that contains G i as an induced subgraph. Since we perform at most c(r) iterations and in each iteration we remove at most ( r 2) γ n edges, the total number of edges removed is bounded by c(r) (r 2) γn < ɛ d n/2. At the end of the process we obtain a graph H with the property that for any graph G i either H contains no copy of G i or it contains at least γ n such copies. Since G was assumed to be ɛfar from Π, and H was obtained from G by removing at most ɛ d n/2 edges, we have that H is ɛ 2 far from Π. Also, since H is obtained from H by removing less than ɛ d n/2 edges, H does not satisfy Π and hence it contains a graph G Hforb Π J r. Now, by the conclusion of the previous paragraph, this means that if G has characteristic vector f 1,..., f c(r) then for every i for which 6
8 f i > 0 we must have that H contains at least γ n copies of G i. Finally, observe that from the definition of the process of obtaining H it follows that H must contain at least this many induced copies of G i. Hence, for every i for which f i > 0 we have g H i γ n. 3.2 Function Ψ Π We now introduce a key notion that we will use to test a hereditary property Π. Note, that the discussion below does not relate to the family of graphs F to which the input instance should belong. Given a family of pairwise nonisomorphic connected graphs {G 1,..., G k } let m({g 1,..., G k }) be the least integer m with the property that the graph that contains m vertex disjoint copies of each of the graphs G i does not satisfy Π. If no such integer m exists, then we set m({g 1,..., G k }) =. For an integer r, let Π r be the family of all sets of pairwise nonisomorphic connected graphs {G 1,..., G k } with the property that all the graphs G i are of size at most r and m({g 1,..., G k }) <. Definition 3.5 For a fixed hereditary property Π we define a function Ψ Π : N N as follows: In case Π r = we set Ψ Π (r) = 0. Ψ Π (r) = max {G 1,...,G k } Π r m({g 1,..., G k }). Note that the above is well defined as for a fixed integer r the set Π r is finite. 3.3 Proof of the main theorem We now formally state and prove the main result of this paper. Theorem 1 Let F be a hereditary and nonexpanding family of graphs. Then every hereditary graph property Π is nonuniformly testable for F with onesided error. Furthermore, Π is uniformly testable with onesided error if ψ Π is computable (or if its approximation is computable, where the quality of the approximation must be independent of the input graph size). Proof : Suppose that G F is ɛfar from Π and consider the subgraph H of G that is obtained via Corollary 3.2. By Lemma 3.4, there is a graph G that does not satisfy Π with the property that all its connected components G i are of size at most r = 2 c/ɛ2 and each of these connected component appears as an induced subgraph of H at least γn times, where γ = ɛd/2 r2. Observe that since each connected component of H is of size at most r, each of these connected components contains at most 2 r copies of each of the connected components G i of G. Therefore, for each G i we have that at least γn/2 r of the connected components of H contains an induced copy of G i. Consider now the set of distinct connected components of G, denoted {G 1,..., G k }. Since G Π we have that m({g 1,..., G k }) < (cf. Section 3.2). Now the definition of Ψ Π guarantees that the graph obtained by taking Ψ Π (r) vertex disjoint copies of each of the graphs G i does not satisfy Π. By the first paragraph of the proof, a randomly chosen vertex belongs to a connected component of H which contains a copy of G i with probability at least γ/2 r. Therefore, by Markov s inequality a randomly chosen sample of size 10 c(r) 2 r Ψ Π (r)/γ will, with probability at least 2/3, contain k Ψ(r) vertices {v i,j } 1 j Ψ Π(r) 1 i k that belong to distinct connected component of H, with the property that for every 1 j Ψ Π (r), the connected component of H to which v i,j belongs, is an induced copy of G i. In particular, the graph that is 7
9 obtained by taking the disjoint union of the connected components to which the vertices v i,j belong does not satisfy Π. Finally, since G does not contain edges connecting vertices from distinct nontrivial connected components of H, we get that any graph that is obtained by taking the union of nontrivial connected components of H is also an induced subgraph of G. Therefore, with probability at least 2/3 the tester will reject G. Also, from the above analysis one can see that we can set s 2 = r = 2 c/ɛ2 and s 1 = 10 c(r) 2 r Ψ Π (r)/γ. 3.4 Discussion When do we need Ψ Π : Notice that the function Ψ Π, defined in Section 3.2 is not necessarily computable. However, we only need this definition in order to obtain a general result on all hereditary properties. Observe, for example, that for any hereditary property Π that is closed under disjoint union 3 we have that Ψ Π (r) = 1. Therefore, in these cases we have a trivial function Ψ. Furthermore, notice that any natural hereditary property, such as those discussed throughout the paper, is closed under disjoint union, therefore for such properties we get uniform testers (for any hereditary family of graphs F). When does Π have a uniform tester: The proof of Theorem 1 shows that when the function Ψ Π is computable then one can design a onesided error uniform tester for Π. Using arguments similar to those used in [8], it can be shown that if the tester is allowed to use the size of the input in order to make its decisions then all hereditary properties have a uniform tester with constant query complexity but with running time that depends on n. Following [8], let us define an oblivious tester as one that has no access to the size of the input when making its decisions. Given ɛ, an oblivious tester computes a number q = Q(ɛ), and then asks an oracle for D(v, q) for all the vertices v S, where S is a random subset of vertices of V (G) of size q (recall that D(v, q) is the neighborhood of v of radius q). Using the answers to these queries the tester should either accept or reject the input. Observe that the algorithm we design in the proof of Theorem 1 is oblivious. Therefore, if Ψ Π is computable, then Π has an oblivious onesided error uniform tester. Let us show that for any hereditary property Π, the computability of Ψ Π is not only sufficient but also necessary, if one wants to design an oblivious onesided error tester for Π. Here is a sketch of the proof. It is easy to see that an oblivious onesided error tester for a hereditary property must accept the input if the graph that is spanned by v S D(v, q) satisfies the property 4. Suppose then that Π can be tested with query complexity Q(ɛ). We claim that in this case Ψ Π (r) Q(1/2 r2 ) and since Q is assumed to be computable, then so does Ψ Π. Indeed, for any {G 1,..., G k } Π r and for any positive integer d, consider a graph consisting of d disjoint copies of each graph G i. Let us think of this graph as consisting of d clusters C j, where each cluster C j contains one copy of each of graphs G 1,..., G k. This graph has degree bounded by r and we claim that for all large enough d, it is 1/2 r2 far from Π. Let us denote by n the number of vertices of the graph and by m the number of vertices in each cluster C i, and observe that m r2 (r 2). Therefore, after n n adding/removing at most 4m edges, we will still have 2m clusters C j which have not changed. Therefore, as m({g 1,..., G k }) < for large enough d, the new graph still does not satisfy Π. We thus conclude that for large enough d, the graph is at least 1/(4mr)far from satisfying Π (and 1/(4mr) 1/2 r2 ). However, since the algorithm must find a graph that does not satisfy Π, it must ask at least m({g 1,..., G k }) queries in 3 That is, if G 1 = (V 1, E 1 ) and G 2 = (V 2, E 2 ) satisfy the property, then so does G 3 = (V 1 V 2, E 1 E 2 ). 4 Suppose the tester rejects an input even though v S D(v, q) satisfies Π. In that case if we were to execute the tester on the graph that is defined as the disjoint union of {D(v, q) : v S} it would have a nonzero probability of rejecting this graph even though it satisfies the property. 8
10 order to succeed on such graphs. Therefore, m({g 1,..., G k }) Q(1/2 r2 ) for any set {G 1,..., G k } Π r and by the definition of Ψ Π this means that Ψ Π Q(1/2 r2 ) as needed. 4 Conclusions In this paper we made a first attempt to give general testability results for graphs belonging to restricted families of graphs. We showed that all hereditary graph properties are (nonuniformly) testable, if the input graph comes from a family of graphs that is hereditary and nonexpanding. Some interesting open questions include. Which properties can be tested for expander graphs? Which properties can be tested in O( n) time for expander graphs? Which properties can be tested for nonexpanding families of graphs when only the average degree of the graph is bounded? Which properties can be tested for directed graphs in sublinear time (in particular, when we can see a directed edge u, v only from vertex u)? References [1] N. Alon, E. Fischer, M. Krivelevich, M. Szegedy. Efficient testing of large graphs. Combinatorica, 20(4): , [2] N. Alon, E. Fischer, I. Newman, and A. Shapira. A combinatorial characterization of the testable graph properties: it s all about regularity. Proceedings of the 38th Annual ACM Symposium on Theory of Computing (STOC), pp , [3] N. Alon, T. Kaufman, M. Krivelevich, and D. Ron. Testing trianglefreeness in general graphs. Proceedings of the 17th Annual ACMSIAM Symposium on Discrete Algorithms (SODA), pp , [4] N. Alon, P. Seymour, and R. Thomas. A separator theorem for graphs with an excluded minor and its applications. Proceedings of the 27th Annual ACM Symposium on Theory of Computing (STOC), pp , [5] N. Alon and A. Shapira. Every monotone graph property is testable. Proceedings of the 37th Annual ACM Symposium on Theory of Computing (STOC), pp , [6] N. Alon and A. Shapira. A characterization of the (natural) graph properties testable with onesided error. Proceedings of the 46th IEEE Symposium on Foundations of Computer Science (FOCS), pp , [7] N. Alon and A. Shapira. Homomorphisms in graph property testing  A survey. Topics in Discrete Mathematics, dedicated to Jarik Nesetril on the occasion of his 60th Birthday, edited by M. Klazar, J. Kratochvil, M. Loebl, J. Matoušek, R. Thomas and P. Valtr, pp [8] N. Alon and A. Shapira. A separation theorem in property testing, submitted. 9
11 [9] M. A. Bender and D. Ron. Testing properties of directed graphs: acyclicity and connectivity. Random Structures and Algorithms, 20(2): , [10] A. Bogdanov, K. Obata, and L. Trevisan. A lower bound for testing 3colorability in boundeddegree graphs. Proceedings of the 43rd IEEE Symposium on Foundations of Computer Science (FOCS), pp , [11] C. Borgs, J. Chayes, L. Lovász, V. T. Sos, B. Szegedy, and K. Vesztergombi. Graph limits and parameter testing. Proceedings of the 38th Annual ACM Symposium on Theory of Computing (STOC), pp , [12] A. Czumaj and C. Sohler. Abstract combinatorial programs and efficient property testers. SIAM Journal on Computing, 34(3): , [13] A. Czumaj and C. Sohler. Sublineartime algorithms Bulletin of the EATCS, 89: 23 47, June [14] E. Fischer. The art of uninformed decisions: A primer to property testing. Bulletin of the EATCS, 75: , October [15] J. R. Gilbert, J. P. Hutchinson, and R. E. Tarjan. A separator theorem for graphs of bounded genus. Journal of Algorithms, 5: , [16] O. Goldreich. Combinatorial property testing (a survey). In P. Pardalos, S. Rajasekaran, and J. Rolim, editors, Proceedings of the DIMACS Workshop on Randomization Methods in Algorithm Design, volume 43 of DIMACS, Series in Discrete Mathetaics and Theoretical Computer Science, pp , American Mathematical Society, Providence, RI, [17] O. Goldreich. Property testing in massive graphs. In J. Abello, P. M. Pardalos, and M. G. C. Resende, editors, Handbook of Massive Data Sets, pp Kluwer Academic Publishers, [18] O. Goldreich. Contemplations on testing graph properties. Manuscript, August [19] O. Goldreich and D. Ron. A sublinear bipartiteness tester for bounded degree graphs. Combinatorica, 19(3): , [20] O. Goldreich and D. Ron. On testing expansion in boundeddegree graphs. Electronic Colloquium on Computational Complexity (ECCC), Report No. 7, [21] O. Goldreich and D. Ron. Property testing in bounded degree graphs. Algorithmica, 32(2): , [22] M. C. Golumbic. Algorithmic Graph Theory and Perfect Graphs. Academic Press, [23] R. Kumar and R. Rubinfeld. Sublinear time algorithms. SIGACT News, 34: 57 67, [24] T. Kaufman, M. Krivelevich, and D. Ron. Tight bounds for testing bipartiteness in general graphs. SIAM Journal on Computing, 33(6): , [25] R. J. Lipton and R. E. Tarjan A separator theorem for planar graphs. SIAM Journal of Applied Mathematics, 36(2): ,
12 [26] L. Lovász and B. Szegedy. Graph limits and testing hereditary graph properties. Technical Report, MSRTR , Microsoft Research, August [27] M. Parnas and D. Ron. Testing the diameter of graphs. Random Structures and Algorithms, 20(2): , [28] J. L. RamirezAlfonsin and B. A. Reed, editors. Perfect Graphs. Wiley and Sons, [29] D. Ron. Property testing. In P. M. Pardalos, S. Rajasekaran, J. Reif, and J. D. P. Rolim, editors, Handobook of Randomized Algorithms, volume II, pp Kluwer Academic Publishers,
A Brief Introduction to Property Testing
A Brief Introduction to Property Testing Oded Goldreich Abstract. This short article provides a brief description of the main issues that underly the study of property testing. It is meant to serve as
More informationCSC2420 Fall 2012: Algorithm Design, Analysis and Theory
CSC2420 Fall 2012: Algorithm Design, Analysis and Theory Allan Borodin November 15, 2012; Lecture 10 1 / 27 Randomized online bipartite matching and the adwords problem. We briefly return to online algorithms
More informationA Sublinear Bipartiteness Tester for Bounded Degree Graphs
A Sublinear Bipartiteness Tester for Bounded Degree Graphs Oded Goldreich Dana Ron February 5, 1998 Abstract We present a sublineartime algorithm for testing whether a bounded degree graph is bipartite
More informationMean RamseyTurán numbers
Mean RamseyTurán numbers Raphael Yuster Department of Mathematics University of Haifa at Oranim Tivon 36006, Israel Abstract A ρmean coloring of a graph is a coloring of the edges such that the average
More informationOn an antiramsey type result
On an antiramsey type result Noga Alon, Hanno Lefmann and Vojtĕch Rödl Abstract We consider antiramsey type results. For a given coloring of the kelement subsets of an nelement set X, where two kelement
More informationA Turán Type Problem Concerning the Powers of the Degrees of a Graph
A Turán Type Problem Concerning the Powers of the Degrees of a Graph Yair Caro and Raphael Yuster Department of Mathematics University of HaifaORANIM, Tivon 36006, Israel. AMS Subject Classification:
More informationCycle transversals in bounded degree graphs
Electronic Notes in Discrete Mathematics 35 (2009) 189 195 www.elsevier.com/locate/endm Cycle transversals in bounded degree graphs M. Groshaus a,2,3 P. Hell b,3 S. Klein c,1,3 L. T. Nogueira d,1,3 F.
More informationHigh degree graphs contain largestar factors
High degree graphs contain largestar factors Dedicated to László Lovász, for his 60th birthday Noga Alon Nicholas Wormald Abstract We show that any finite simple graph with minimum degree d contains a
More informationOn the kpath cover problem for cacti
On the kpath cover problem for cacti Zemin Jin and Xueliang Li Center for Combinatorics and LPMC Nankai University Tianjin 300071, P.R. China zeminjin@eyou.com, x.li@eyou.com Abstract In this paper we
More informationCOMBINATORIAL PROPERTIES OF THE HIGMANSIMS GRAPH. 1. Introduction
COMBINATORIAL PROPERTIES OF THE HIGMANSIMS GRAPH ZACHARY ABEL 1. Introduction In this survey we discuss properties of the HigmanSims graph, which has 100 vertices, 1100 edges, and is 22 regular. In fact
More informationON INDUCED SUBGRAPHS WITH ALL DEGREES ODD. 1. Introduction
ON INDUCED SUBGRAPHS WITH ALL DEGREES ODD A.D. SCOTT Abstract. Gallai proved that the vertex set of any graph can be partitioned into two sets, each inducing a subgraph with all degrees even. We prove
More informationShort Cycles make Whard problems hard: FPT algorithms for Whard Problems in Graphs with no short Cycles
Short Cycles make Whard problems hard: FPT algorithms for Whard Problems in Graphs with no short Cycles Venkatesh Raman and Saket Saurabh The Institute of Mathematical Sciences, Chennai 600 113. {vraman
More informationPh.D. Thesis. Judit NagyGyörgy. Supervisor: Péter Hajnal Associate Professor
Online algorithms for combinatorial problems Ph.D. Thesis by Judit NagyGyörgy Supervisor: Péter Hajnal Associate Professor Doctoral School in Mathematics and Computer Science University of Szeged Bolyai
More informationRandom graphs with a given degree sequence
Sourav Chatterjee (NYU) Persi Diaconis (Stanford) Allan Sly (Microsoft) Let G be an undirected simple graph on n vertices. Let d 1,..., d n be the degrees of the vertices of G arranged in descending order.
More informationLinearity and Group Homomorphism Testing / Testing Hadamard Codes
Title: Linearity and Group Homomorphism Testing / Testing Hadamard Codes Name: Sofya Raskhodnikova 1, Ronitt Rubinfeld 2 Affil./Addr. 1: Pennsylvania State University Affil./Addr. 2: Keywords: SumOriWork:
More informationRamsey numbers for bipartite graphs with small bandwidth
Ramsey numbers for bipartite graphs with small bandwidth Guilherme O. Mota 1,, Gábor N. Sárközy 2,, Mathias Schacht 3,, and Anusch Taraz 4, 1 Instituto de Matemática e Estatística, Universidade de São
More informationLabeling outerplanar graphs with maximum degree three
Labeling outerplanar graphs with maximum degree three Xiangwen Li 1 and Sanming Zhou 2 1 Department of Mathematics Huazhong Normal University, Wuhan 430079, China 2 Department of Mathematics and Statistics
More informationLarge induced subgraphs with all degrees odd
Large induced subgraphs with all degrees odd A.D. Scott Department of Pure Mathematics and Mathematical Statistics, University of Cambridge, England Abstract: We prove that every connected graph of order
More informationFinding and counting given length cycles
Finding and counting given length cycles Noga Alon Raphael Yuster Uri Zwick Abstract We present an assortment of methods for finding and counting simple cycles of a given length in directed and undirected
More informationA 2factor in which each cycle has long length in clawfree graphs
A factor in which each cycle has long length in clawfree graphs Roman Čada Shuya Chiba Kiyoshi Yoshimoto 3 Department of Mathematics University of West Bohemia and Institute of Theoretical Computer Science
More informationDefinition 11.1. Given a graph G on n vertices, we define the following quantities:
Lecture 11 The Lovász ϑ Function 11.1 Perfect graphs We begin with some background on perfect graphs. graphs. First, we define some quantities on Definition 11.1. Given a graph G on n vertices, we define
More information2.3 Scheduling jobs on identical parallel machines
2.3 Scheduling jobs on identical parallel machines There are jobs to be processed, and there are identical machines (running in parallel) to which each job may be assigned Each job = 1,,, must be processed
More informationCycles and cliqueminors in expanders
Cycles and cliqueminors in expanders Benny Sudakov UCLA and Princeton University Expanders Definition: The vertex boundary of a subset X of a graph G: X = { all vertices in G\X with at least one neighbor
More informationCompetitive Analysis of On line Randomized Call Control in Cellular Networks
Competitive Analysis of On line Randomized Call Control in Cellular Networks Ioannis Caragiannis Christos Kaklamanis Evi Papaioannou Abstract In this paper we address an important communication issue arising
More informationApproximated Distributed Minimum Vertex Cover Algorithms for Bounded Degree Graphs
Approximated Distributed Minimum Vertex Cover Algorithms for Bounded Degree Graphs Yong Zhang 1.2, Francis Y.L. Chin 2, and HingFung Ting 2 1 College of Mathematics and Computer Science, Hebei University,
More informationGRAPH THEORY LECTURE 4: TREES
GRAPH THEORY LECTURE 4: TREES Abstract. 3.1 presents some standard characterizations and properties of trees. 3.2 presents several different types of trees. 3.7 develops a counting method based on a bijection
More informationClique coloring B 1 EPG graphs
Clique coloring B 1 EPG graphs Flavia Bonomo a,c, María Pía Mazzoleni b,c, and Maya Stein d a Departamento de Computación, FCENUBA, Buenos Aires, Argentina. b Departamento de Matemática, FCEUNLP, La
More informationGraphs without proper subgraphs of minimum degree 3 and short cycles
Graphs without proper subgraphs of minimum degree 3 and short cycles Lothar Narins, Alexey Pokrovskiy, Tibor Szabó Department of Mathematics, Freie Universität, Berlin, Germany. August 22, 2014 Abstract
More informationAnalysis of Algorithms, I
Analysis of Algorithms, I CSOR W4231.002 Eleni Drinea Computer Science Department Columbia University Thursday, February 26, 2015 Outline 1 Recap 2 Representing graphs 3 Breadthfirst search (BFS) 4 Applications
More informationThe degree, size and chromatic index of a uniform hypergraph
The degree, size and chromatic index of a uniform hypergraph Noga Alon Jeong Han Kim Abstract Let H be a kuniform hypergraph in which no two edges share more than t common vertices, and let D denote the
More informationGeneralized Induced Factor Problems
Egerváry Research Group on Combinatorial Optimization Technical reports TR200207. Published by the Egrerváry Research Group, Pázmány P. sétány 1/C, H 1117, Budapest, Hungary. Web site: www.cs.elte.hu/egres.
More informationarxiv: v2 [math.co] 30 Nov 2015
PLANAR GRAPH IS ON FIRE PRZEMYSŁAW GORDINOWICZ arxiv:1311.1158v [math.co] 30 Nov 015 Abstract. Let G be any connected graph on n vertices, n. Let k be any positive integer. Suppose that a fire breaks out
More informationLecture 15 An Arithmetic Circuit Lowerbound and Flows in Graphs
CSE599s: Extremal Combinatorics November 21, 2011 Lecture 15 An Arithmetic Circuit Lowerbound and Flows in Graphs Lecturer: Anup Rao 1 An Arithmetic Circuit Lower Bound An arithmetic circuit is just like
More informationWeek 5 Integral Polyhedra
Week 5 Integral Polyhedra We have seen some examples 1 of linear programming formulation that are integral, meaning that every basic feasible solution is an integral vector. This week we develop a theory
More informationOdd induced subgraphs in graphs of maximum degree three
Odd induced subgraphs in graphs of maximum degree three David M. Berman, Hong Wang, and Larry Wargo Department of Mathematics University of New Orleans New Orleans, Louisiana, USA 70148 Abstract A longstanding
More informationDistance Approximation in BoundedDegree and General Sparse Graphs
Distance Approximation in BoundedDegree and General Sparse Graphs Sharon Marko Dana Ron June 5, 2006 Abstract We address the problem of approximating the distance of bounded degree and general sparse
More informationCMSC 858T: Randomized Algorithms Spring 2003 Handout 8: The Local Lemma
CMSC 858T: Randomized Algorithms Spring 2003 Handout 8: The Local Lemma Please Note: The references at the end are given for extra reading if you are interested in exploring these ideas further. You are
More informationThe Open University s repository of research publications and other research outputs
Open Research Online The Open University s repository of research publications and other research outputs The degreediameter problem for circulant graphs of degree 8 and 9 Journal Article How to cite:
More informationMinimum Bisection is NPhard on Unit Disk Graphs
Minimum Bisection is NPhard on Unit Disk Graphs Josep Díaz 1 and George B. Mertzios 2 1 Departament de Llenguatges i Sistemes Informátics, Universitat Politécnica de Catalunya, Spain. 2 School of Engineering
More informationAnalysis of Approximation Algorithms for kset Cover using FactorRevealing Linear Programs
Analysis of Approximation Algorithms for kset Cover using FactorRevealing Linear Programs Stavros Athanassopoulos, Ioannis Caragiannis, and Christos Kaklamanis Research Academic Computer Technology Institute
More informationTHe area of sublineartime algorithms is a new rapidly emerging area of computer science. It
Sublineartime algorithms Artur Czumaj Christian Sohler Abstract In this paper we survey recent advances in the area of sublineartime algorithms. Introduction THe area of sublineartime algorithms is
More informationAll trees contain a large induced subgraph having all degrees 1 (mod k)
All trees contain a large induced subgraph having all degrees 1 (mod k) David M. Berman, A.J. Radcliffe, A.D. Scott, Hong Wang, and Larry Wargo *Department of Mathematics University of New Orleans New
More informationAbout the inverse football pool problem for 9 games 1
Seventh International Workshop on Optimal Codes and Related Topics September 61, 013, Albena, Bulgaria pp. 15133 About the inverse football pool problem for 9 games 1 Emil Kolev Tsonka Baicheva Institute
More informationThe chromatic spectrum of mixed hypergraphs
The chromatic spectrum of mixed hypergraphs Tao Jiang, Dhruv Mubayi, Zsolt Tuza, Vitaly Voloshin, Douglas B. West March 30, 2003 Abstract A mixed hypergraph is a triple H = (X, C, D), where X is the vertex
More informationCOUNTING INDEPENDENT SETS IN SOME CLASSES OF (ALMOST) REGULAR GRAPHS
COUNTING INDEPENDENT SETS IN SOME CLASSES OF (ALMOST) REGULAR GRAPHS Alexander Burstein Department of Mathematics Howard University Washington, DC 259, USA aburstein@howard.edu Sergey Kitaev Mathematics
More informationNimble Algorithms for Cloud Computing. Ravi Kannan, Santosh Vempala and David Woodruff
Nimble Algorithms for Cloud Computing Ravi Kannan, Santosh Vempala and David Woodruff Cloud computing Data is distributed arbitrarily on many servers Parallel algorithms: time Streaming algorithms: sublinear
More informationFairness in Routing and Load Balancing
Fairness in Routing and Load Balancing Jon Kleinberg Yuval Rabani Éva Tardos Abstract We consider the issue of network routing subject to explicit fairness conditions. The optimization of fairness criteria
More informationTesting LowDegree Polynomials over
Testing LowDegree Polynomials over Noga Alon, Tali Kaufman, Michael Krivelevich, Simon Litsyn, and Dana Ron Department of Mathematics, Tel Aviv University, Tel Aviv 69978, Israel and Institute for Advanced
More informationDiscrete Mathematics & Mathematical Reasoning Chapter 10: Graphs
Discrete Mathematics & Mathematical Reasoning Chapter 10: Graphs Kousha Etessami U. of Edinburgh, UK Kousha Etessami (U. of Edinburgh, UK) Discrete Mathematics (Chapter 6) 1 / 13 Overview Graphs and Graph
More informationExponential time algorithms for graph coloring
Exponential time algorithms for graph coloring Uriel Feige Lecture notes, March 14, 2011 1 Introduction Let [n] denote the set {1,..., k}. A klabeling of vertices of a graph G(V, E) is a function V [k].
More informationThe Independence Number in Graphs of Maximum Degree Three
The Independence Number in Graphs of Maximum Degree Three Jochen Harant 1 Michael A. Henning 2 Dieter Rautenbach 1 and Ingo Schiermeyer 3 1 Institut für Mathematik, TU Ilmenau, Postfach 100565, D98684
More informationPartitioning edgecoloured complete graphs into monochromatic cycles and paths
arxiv:1205.5492v1 [math.co] 24 May 2012 Partitioning edgecoloured complete graphs into monochromatic cycles and paths Alexey Pokrovskiy Departement of Mathematics, London School of Economics and Political
More informationTenacity and rupture degree of permutation graphs of complete bipartite graphs
Tenacity and rupture degree of permutation graphs of complete bipartite graphs Fengwei Li, Qingfang Ye and Xueliang Li Department of mathematics, Shaoxing University, Shaoxing Zhejiang 312000, P.R. China
More informationThe positive minimum degree game on sparse graphs
The positive minimum degree game on sparse graphs József Balogh Department of Mathematical Sciences University of Illinois, USA jobal@math.uiuc.edu András Pluhár Department of Computer Science University
More information136 CHAPTER 4. INDUCTION, GRAPHS AND TREES
136 TER 4. INDUCTION, GRHS ND TREES 4.3 Graphs In this chapter we introduce a fundamental structural idea of discrete mathematics, that of a graph. Many situations in the applications of discrete mathematics
More informationTreerepresentation of set families and applications to combinatorial decompositions
Treerepresentation of set families and applications to combinatorial decompositions BinhMinh BuiXuan a, Michel Habib b Michaël Rao c a Department of Informatics, University of Bergen, Norway. buixuan@ii.uib.no
More informationStrictly Localized Construction of Planar BoundedDegree Spanners of Unit Disk Graphs
Strictly Localized Construction of Planar BoundedDegree Spanners of Unit Disk Graphs Iyad A. Kanj Ljubomir Perković Ge Xia Abstract We describe a new distributed and strictlylocalized approach for constructing
More informationLecture 3: Linear Programming Relaxations and Rounding
Lecture 3: Linear Programming Relaxations and Rounding 1 Approximation Algorithms and Linear Relaxations For the time being, suppose we have a minimization problem. Many times, the problem at hand can
More informationEmbedding nearlyspanning bounded degree trees
Embedding nearlyspanning bounded degree trees Noga Alon Michael Krivelevich Benny Sudakov Abstract We derive a sufficient condition for a sparse graph G on n vertices to contain a copy of a tree T of
More informationOffline sorting buffers on Line
Offline sorting buffers on Line Rohit Khandekar 1 and Vinayaka Pandit 2 1 University of Waterloo, ON, Canada. email: rkhandekar@gmail.com 2 IBM India Research Lab, New Delhi. email: pvinayak@in.ibm.com
More informationmost 4 Mirka Miller 1,2, Guillermo PinedaVillavicencio 3, The University of Newcastle Callaghan, NSW 2308, Australia University of West Bohemia
Complete catalogue of graphs of maimum degree 3 and defect at most 4 Mirka Miller 1,2, Guillermo PinedaVillavicencio 3, 1 School of Electrical Engineering and Computer Science The University of Newcastle
More informationOn Integer Additive SetIndexers of Graphs
On Integer Additive SetIndexers of Graphs arxiv:1312.7672v4 [math.co] 2 Mar 2014 N K Sudev and K A Germina Abstract A setindexer of a graph G is an injective setvalued function f : V (G) 2 X such that
More informationOn the independence number of graphs with maximum degree 3
On the independence number of graphs with maximum degree 3 Iyad A. Kanj Fenghui Zhang Abstract Let G be an undirected graph with maximum degree at most 3 such that G does not contain any of the three graphs
More informationStrong Ramsey Games: Drawing on an infinite board
Strong Ramsey Games: Drawing on an infinite board Dan Hefetz Christopher Kusch Lothar Narins Alexey Pokrovskiy Clément Requilé Amir Sarid arxiv:1605.05443v2 [math.co] 25 May 2016 May 26, 2016 Abstract
More informationSHARP BOUNDS FOR THE SUM OF THE SQUARES OF THE DEGREES OF A GRAPH
31 Kragujevac J. Math. 25 (2003) 31 49. SHARP BOUNDS FOR THE SUM OF THE SQUARES OF THE DEGREES OF A GRAPH Kinkar Ch. Das Department of Mathematics, Indian Institute of Technology, Kharagpur 721302, W.B.,
More informationDegree Hypergroupoids Associated with Hypergraphs
Filomat 8:1 (014), 119 19 DOI 10.98/FIL1401119F Published by Faculty of Sciences and Mathematics, University of Niš, Serbia Available at: http://www.pmf.ni.ac.rs/filomat Degree Hypergroupoids Associated
More informationA REMARK ON ALMOST MOORE DIGRAPHS OF DEGREE THREE. 1. Introduction and Preliminaries
Acta Math. Univ. Comenianae Vol. LXVI, 2(1997), pp. 285 291 285 A REMARK ON ALMOST MOORE DIGRAPHS OF DEGREE THREE E. T. BASKORO, M. MILLER and J. ŠIRÁŇ Abstract. It is well known that Moore digraphs do
More informationINDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS
INDISTINGUISHABILITY OF ABSOLUTELY CONTINUOUS AND SINGULAR DISTRIBUTIONS STEVEN P. LALLEY AND ANDREW NOBEL Abstract. It is shown that there are no consistent decision rules for the hypothesis testing problem
More informationMinimum degree condition forcing complete graph immersion
Minimum degree condition forcing complete graph immersion Matt DeVos Department of Mathematics Simon Fraser University Burnaby, B.C. V5A 1S6 Jacob Fox Department of Mathematics MIT Cambridge, MA 02139
More informationThe Conference Call Search Problem in Wireless Networks
The Conference Call Search Problem in Wireless Networks Leah Epstein 1, and Asaf Levin 2 1 Department of Mathematics, University of Haifa, 31905 Haifa, Israel. lea@math.haifa.ac.il 2 Department of Statistics,
More information3. Linear Programming and Polyhedral Combinatorics
Massachusetts Institute of Technology Handout 6 18.433: Combinatorial Optimization February 20th, 2009 Michel X. Goemans 3. Linear Programming and Polyhedral Combinatorics Summary of what was seen in the
More informationWhen is a graph planar?
When is a graph planar? Theorem(Euler, 1758) If a plane multigraph G with k components has n vertices, e edges, and f faces, then n e + f = 1 + k. Corollary If G is a simple, planar graph with n(g) 3,
More informationBest Monotone Degree Bounds for Various Graph Parameters
Best Monotone Degree Bounds for Various Graph Parameters D. Bauer Department of Mathematical Sciences Stevens Institute of Technology Hoboken, NJ 07030 S. L. Hakimi Department of Electrical and Computer
More informationCombinatorial 5/6approximation of Max Cut in graphs of maximum degree 3
Combinatorial 5/6approximation of Max Cut in graphs of maximum degree 3 Cristina Bazgan a and Zsolt Tuza b,c,d a LAMSADE, Université ParisDauphine, Place du Marechal de Lattre de Tassigny, F75775 Paris
More informationError Compensation in Leaf Power Problems
Error Compensation in Leaf Power Problems Michael Dom Jiong Guo Falk Hüffner Rolf Niedermeier April 6, 2005 Abstract The kleaf Power recognition problem is a particular case of graph power problems: For
More informationOn end degrees and infinite cycles in locally finite graphs
On end degrees and infinite cycles in locally finite graphs Henning Bruhn Maya Stein Abstract We introduce a natural extension of the vertex degree to ends. For the cycle space C(G) as proposed by Diestel
More informationON DEGREES IN THE HASSE DIAGRAM OF THE STRONG BRUHAT ORDER
Séminaire Lotharingien de Combinatoire 53 (2006), Article B53g ON DEGREES IN THE HASSE DIAGRAM OF THE STRONG BRUHAT ORDER RON M. ADIN AND YUVAL ROICHMAN Abstract. For a permutation π in the symmetric group
More informationFactoring & Primality
Factoring & Primality Lecturer: Dimitris Papadopoulos In this lecture we will discuss the problem of integer factorization and primality testing, two problems that have been the focus of a great amount
More informationRemoving Even Crossings
EuroComb 2005 DMTCS proc. AE, 2005, 105 110 Removing Even Crossings Michael J. Pelsmajer 1, Marcus Schaefer 2 and Daniel Štefankovič 2 1 Department of Applied Mathematics, Illinois Institute of Technology,
More informationOn three zerosum Ramseytype problems
On three zerosum Ramseytype problems Noga Alon Department of Mathematics Raymond and Beverly Sackler Faculty of Exact Sciences Tel Aviv University, Tel Aviv, Israel and Yair Caro Department of Mathematics
More informationPolynomial Degree and Lower Bounds in Quantum Complexity: Collision and Element Distinctness with Small Range
THEORY OF COMPUTING, Volume 1 (2005), pp. 37 46 http://theoryofcomputing.org Polynomial Degree and Lower Bounds in Quantum Complexity: Collision and Element Distinctness with Small Range Andris Ambainis
More informationThe average distances in random graphs with given expected degrees
Classification: Physical Sciences, Mathematics The average distances in random graphs with given expected degrees by Fan Chung 1 and Linyuan Lu Department of Mathematics University of California at San
More informationAn Approximation Algorithm for Bounded Degree Deletion
An Approximation Algorithm for Bounded Degree Deletion Tomáš Ebenlendr Petr Kolman Jiří Sgall Abstract Bounded Degree Deletion is the following generalization of Vertex Cover. Given an undirected graph
More informationGraph theoretic techniques in the analysis of uniquely localizable sensor networks
Graph theoretic techniques in the analysis of uniquely localizable sensor networks Bill Jackson 1 and Tibor Jordán 2 ABSTRACT In the network localization problem the goal is to determine the location of
More informationAsking Hard Graph Questions. Paul Burkhardt. February 3, 2014
Beyond Watson: Predictive Analytics and Big Data U.S. National Security Agency Research Directorate  R6 Technical Report February 3, 2014 300 years before Watson there was Euler! The first (Jeopardy!)
More informationSHORT CYCLE COVERS OF GRAPHS WITH MINIMUM DEGREE THREE
SHOT YLE OVES OF PHS WITH MINIMUM DEEE THEE TOMÁŠ KISE, DNIEL KÁL, END LIDIKÝ, PVEL NEJEDLÝ OET ŠÁML, ND bstract. The Shortest ycle over onjecture of lon and Tarsi asserts that the edges of every bridgeless
More information1 The Line vs Point Test
6.875 PCP and Hardness of Approximation MIT, Fall 2010 Lecture 5: Low Degree Testing Lecturer: Dana Moshkovitz Scribe: Gregory Minton and Dana Moshkovitz Having seen a probabilistic verifier for linearity
More informationUPPER BOUNDS ON THE L(2, 1)LABELING NUMBER OF GRAPHS WITH MAXIMUM DEGREE
UPPER BOUNDS ON THE L(2, 1)LABELING NUMBER OF GRAPHS WITH MAXIMUM DEGREE ANDREW LUM ADVISOR: DAVID GUICHARD ABSTRACT. L(2,1)labeling was first defined by Jerrold Griggs [Gr, 1992] as a way to use graphs
More informationA simpler and better derandomization of an approximation algorithm for Single Source RentorBuy
A simpler and better derandomization of an approximation algorithm for Single Source RentorBuy David P. Williamson Anke van Zuylen School of Operations Research and Industrial Engineering, Cornell University,
More informationDiscrete Applied Mathematics. The firefighter problem with more than one firefighter on trees
Discrete Applied Mathematics 161 (2013) 899 908 Contents lists available at SciVerse ScienceDirect Discrete Applied Mathematics journal homepage: www.elsevier.com/locate/dam The firefighter problem with
More informationSemiMatchings for Bipartite Graphs and Load Balancing
SemiMatchings for Bipartite Graphs and Load Balancing Nicholas J. A. Harvey Richard E. Ladner László Lovász Tami Tamir Abstract We consider the problem of fairly matching the lefthand vertices of a bipartite
More informationCHAPTER II THE LIMIT OF A SEQUENCE OF NUMBERS DEFINITION OF THE NUMBER e.
CHAPTER II THE LIMIT OF A SEQUENCE OF NUMBERS DEFINITION OF THE NUMBER e. This chapter contains the beginnings of the most important, and probably the most subtle, notion in mathematical analysis, i.e.,
More informationSOLUTIONS TO ASSIGNMENT 1 MATH 576
SOLUTIONS TO ASSIGNMENT 1 MATH 576 SOLUTIONS BY OLIVIER MARTIN 13 #5. Let T be the topology generated by A on X. We want to show T = J B J where B is the set of all topologies J on X with A J. This amounts
More informationEvery tree contains a large induced subgraph with all degrees odd
Every tree contains a large induced subgraph with all degrees odd A.J. Radcliffe Carnegie Mellon University, Pittsburgh, PA A.D. Scott Department of Pure Mathematics and Mathematical Statistics University
More informationOn onefactorizations of replacement products
Filomat 27:1 (2013), 57 63 DOI 10.2298/FIL1301057A Published by Faculty of Sciences and Mathematics, University of Niš, Serbia Available at: http://www.pmf.ni.ac.rs/filomat On onefactorizations of replacement
More informationUniform Multicommodity Flow through the Complete Graph with Random Edgecapacities
Combinatorics, Probability and Computing (2005) 00, 000 000. c 2005 Cambridge University Press DOI: 10.1017/S0000000000000000 Printed in the United Kingdom Uniform Multicommodity Flow through the Complete
More informationDivideandConquer. Three main steps : 1. divide; 2. conquer; 3. merge.
DivideandConquer Three main steps : 1. divide; 2. conquer; 3. merge. 1 Let I denote the (sub)problem instance and S be its solution. The divideandconquer strategy can be described as follows. Procedure
More informationCS 598CSC: Combinatorial Optimization Lecture date: 2/4/2010
CS 598CSC: Combinatorial Optimization Lecture date: /4/010 Instructor: Chandra Chekuri Scribe: David Morrison GomoryHu Trees (The work in this section closely follows [3]) Let G = (V, E) be an undirected
More informationTypes of Degrees in Bipolar Fuzzy Graphs
pplied Mathematical Sciences, Vol. 7, 2013, no. 98, 48574866 HIKRI Ltd, www.mhikari.com http://dx.doi.org/10.12988/ams.2013.37389 Types of Degrees in Bipolar Fuzzy Graphs Basheer hamed Mohideen Department
More informationLecture 1: Course overview, circuits, and formulas
Lecture 1: Course overview, circuits, and formulas Topics in Complexity Theory and Pseudorandomness (Spring 2013) Rutgers University Swastik Kopparty Scribes: John Kim, Ben Lund 1 Course Information Swastik
More information